metadata

annotations_creators:
  - expert-generated
language_creators:
  - found
language:
  - id
multilinguality:
  - monolingual
size_categories:
  - n<1K
task_categories:
  - feature-extraction
  - sentence-similarity
pretty_name: BPS Pairs and Triplets Dataset
tags:
  - pairs
  - triplets
  - information-retrieval
configs:
  - config_name: pairs-class
    data_files:
      - split: train
        path: pairs-class/train/*
  - config_name: triplet-all
    data_files:
      - split: train
        path: triplet-all/train/*
dataset_info:
  - config_name: pairs-class
    features:
      - name: query
        dtype: string
      - name: doc
        dtype: string
      - name: label
        dtype: int64
    splits:
      - name: train
        num_bytes: 0
        num_examples: 0
    download_size: 0
    dataset_size: 0
  - config_name: triplet-all
    features:
      - name: anchor
        dtype: string
      - name: positive
        dtype: string
      - name: negative
        dtype: string
    splits:
      - name: train
        num_bytes: 0
        num_examples: 0
    download_size: 0
    dataset_size: 0

BPS Pairs and Triplets Dataset

This dataset contains two configurations for training information retrieval models:

pairs-class: Pairs of queries and documents with relevance labels (currently empty).
- Split: train
- Number of examples: 0
- Each example consists of:
  - query: The query text.
  - doc: The document text.
  - label: Relevance label (e.g., 0 or 1).
triplet-all: Triplets of queries with positive and negative examples (currently empty).
- Split: train
- Number of examples: 0
- Each example consists of:
  - anchor: The query text.
  - positive: A positive (relevant) document text.
  - negative: A negative (irrelevant) document text.